Action goals and the praxis network: an fMRI study

Abstract

The praxis representation network (PRN) of the left cerebral hemisphere is typically linked to the control of functional interactions with familiar tools. Surprisingly, little is known about the PRN engagement in planning and execution of tool-directed actions motivated by non-functional but purposeful action goals. Here we used functional neuroimaging to perform both univariate and multi-voxel pattern analyses (MVPA) in 20 right-handed participants who planned and later executed, with their dominant and non-dominant hands, disparate grasps of tools for different goals, including: (1) planning simple vs. demanding functional grasps of conveniently vs. inconveniently oriented tools with an intention to immediately use them, (2) planning simple-but non-functional-grasps of inconveniently oriented tools with a goal to pass them to a different person, (3) planning reaching movements directed at such tools with an intention to move/push them with the back of the hand, and (4) pantomimed execution of the earlier planned tasks. While PRN contributed to the studied interactions with tools, the engagement of its critical nodes, and/or complementary right hemisphere processing, was differently modulated by task type. E.g., planning non-functional/structural grasp-to-pass movements of inconveniently oriented tools, regardless of the hand, invoked the left parietal and prefrontal nodes significantly more than simple, non-demanding functional grasps. MVPA corroborated decoding capabilities of critical PRN areas and some of their right hemisphere counterparts. Our findings shed new lights on how performance of disparate action goals influences the extraction of object affordances, and how or to what extent it modulates the neural activity within the parieto-frontal brain networks.

Keywords: Affordances; Grasp planning; Hand–tool interactions; Motor cognition; Multi-voxel pattern analysis; Tool grasping.

Fig. 1

Stimuli, conditions and trial structures. A Examples of stimuli used in the main study. Top panel: tools presented at 45, 0, and 315 degrees. Bottom panel: tools presented at 135, 180, and 225 degrees. B Four study conditions resulting from the combination of three different action goals (indicated by goal cues) and two sets of stimulus orientations. C Trial structure and timing of the main experiment, using an event-related design. D Trial structure and timing of the visual tool use localizer task with a block design. E Trial structure and timing of the background study on structure-based (i.e., grasping to displace or move) vs. function-based pantomimed grasping of tools, utilizing a block design

Fig. 2

Brain areas showing significant increases of neural activity during critical localizer tasks, and background comparisons from the main experiment, involving the planning of tool-directed grasps compared to the planning of the reach-and-move task. A Significantly greater neural activity observed for pantomimed grasp-to-use vs. grasp-to-displace task, shown in warm colors, and its inverse contrast shown in cold colors, collapsed across the dominant right and non-dominant left hand. B Neural activity in pantomimed tool use vs. manually simulated animal movements (shown in warm colors), and its inverse contrast (shown in cold colors), collapsed across hands. C Planning of tool-directed grasp pantomimes with the right hand. D Planning of tool-directed grasp pantomimes with the left hand. Both in (C) and (D), the obtained neural activity was averaged across three different study conditions, involving difficult grasp-to-use, easy grasp-to-use, and grasp-to-pass tasks. E Overlays of neural activity for the three tasks, contrasted separately with the reach-and-move task, but collapsed across the two hands. The obtained clusters, and their most representative slices in panel C and D, were thresholded at least at Z > 3.1, and a corrected cluster significance threshold of p = 0.05. Volumetric surface renderings were obtained by means of trilinear interpolation, and their projection onto mid-thickness inflated, and flat surfaces of the connectome workbench atlas, and subsequently demarcated with borders of multi-modal parcellations implemented in this software. The labels of the involved areas can be found on flat maps, and more detailed descriptions of the obtained effects can be found in the main text

Fig. 3

Brain areas showing significant increases of neural activity during planning different tool-directed grasps, contingent on action goal and tool orientation. Brain areas with significantly greater increases for the planning of demanding grasp-to-use, as compared to grasp-to-pass tasks with (A) the right hand, and (B) the left hand. Areas showing significantly greater increases of neural activity during the planning of grasp-to-pass as compared to easy grasp-to-use tasks with (C) the right hand, and (D) the left hand. Areas with significantly greater increases of neural activity for the planning of demanding as compared to easy grasp-to-use tasks, with (E) the right hand, and (F) the left hand

Fig. 4

Brain areas showing significant decodings of planning grasp-to-use (GTU) and grasp-to-pass (GTP) tasks. A The demanding GTU (dGTU) and GTP tasks decoded for the right hand. B The dGTU and GTP tasks decoded for the left hand. C The easy GTU (eGTU) and GTP tasks decoded for the right hand. D The eGTU and GTP tasks decoded for the left hand. E–J The dGTU, GTP, and eGTU tasks decoded in the context of the reach-and-move (RAM) task, both for the right and left hand. Borders in panels A–D were displayed in areas where significant decoding accuracies were obtained. A constant set of parietal borders was used throughout panels E–J (except for cases with no decoding capabilities) as these borders correspond to the outcomes from the contrast involving planning reach-and-move action from the univariate analysis

Fig. 5

The results of ROI analyses for the planning phase. A–G Mean percent signal change within each ROI is plotted relative to the resting baseline for the following Tasks: demanding grasp-to-use (dGTU) task, easy grasp-to-use (eGTU) task, grasp-to-pass (GTP) task, and reach-and-move (RAM) task. (H) The overview of ROI locations depicted on the Connectome Workbench template brain. Asterisks indicate differences with Bonferroni-corrected p values of at least 0.05 (*), 0.01 (**), or 0.001 (***)